Original Article

Genes and Immunity (2008) 9, 349–357; doi:10.1038/gene.2008.24; published online 17 April 2008

Comprehensive association study of genetic variants in the IL-1 gene family in systemic juvenile idiopathic arthritis

C J W Stock1, E M Ogilvie1, J M Samuel1, M Fife1, C M Lewis2 and P Woo1

  1. 1Centre for Paediatric and Adolescent Rheumatology, Windeyer Institute for Medical Sciences, University College London, London, UK
  2. 2Guy's, Kings and St Thomas' School of Medicine, London, UK

Correspondence: Professor P Woo, Centre for Paediatric and Adolescent Rheumatology, 3rd Floor, Windeyer Institute for Medical Sciences, University College London, 46 Cleveland Street, London W1T 4JF, UK. E-mail: p.woo@ucl.ac.uk

Received 19 December 2007; Revised 25 February 2008; Accepted 25 February 2008; Published online 17 April 2008.

Top

Abstract

Patients with systemic juvenile idiopathic arthritis (sJIA) have a characteristic daily spiking fever and elevated levels of inflammatory cytokines. Members of the interleukin-1 (IL-1) gene family have been implicated in various inflammatory and autoimmune diseases, and treatment with the IL-1 receptor antagonist, Anakinra, shows remarkable improvement in some patients. This work describes the most comprehensive investigation to date of the involvement of the IL-1 gene family in sJIA. A two-stage case–control association study was performed to investigate the two clusters of IL-1 family genes using a tagging single nucleotide polymorphism (SNP) approach. Genotyping data of 130 sJIA patients and 151 controls from stage 1 highlighted eight SNPs in the IL1 ligand cluster region and two SNPs in the IL1 receptor cluster region as showing a significant frequency difference between the populations. These 10 SNPs were typed in an additional 105 sJIA patients and 184 controls in stage 2. Meta-analysis of the genotypes from both stages showed that three IL1 ligand cluster SNPs (rs6712572, rs2071374 and rs1688075) and one IL1 receptor cluster SNP (rs12712122) show evidence of significant association with sJIA. These results indicate that there may be aberrant control of the activity of the IL-1 family in sJIA patients causing the increased susceptibility to the disease.

Keywords:

sJIA, IL-1, genetic association

Top

Introduction

Juvenile idiopathic arthritis (JIA) is a clinically heterogeneous disease with seven subtypes.1 It is thought that this clinical diversity reflects a difference in the underlying pathology and genetic heterogeneity. Despite a spectrum of disease activity, systemic JIA (sJIA) can be the most severe of the subtypes, and is potentially fatal. There are currently no disease-specific markers for sJIA and diagnosis may be difficult as exclusion of other diagnoses by an extensive investigative work up is essential.2

There is much evidence that patients with sJIA have an altered cytokine profile, a possible cause for pathogenic chronic inflammation, when compared to patients with other JIA subtypes and to healthy individuals.2 A number of cytokine candidate gene associations have been found that are specific to the systemic subtype of JIA. This includes an association with interleukin-6 (IL-6) gene polymorphisms3, 4 and a macrophage migration inhibitory factor gene polymorphism,5 which may also be a predictor of poorer outcome after intra-articular injections.6 The most recent association to be shown with sJIA is with the anti-inflammatory cytokines IL-10 and IL-20, including an association with a low-expressing IL10 genotype.7

IL-1 is a pro-inflammatory cytokine that affects nearly every cell type, often working in concert with tumor necrosis factor (TNF). It can also upregulate host defences and function as one of the key cytokines mediating the innate immunity response.8 Located together on a locus on chromosome 2q139, 10 are the genes for the two naturally occurring forms of IL-1: IL-1α and IL-1β, and for the naturally occurring IL-1 receptor antagonist, IL-1Ra. Also within the same locus are the six recently discovered genes belonging to the IL-1 family, named IL-1F5–IL-1F1011, 12, 13 (Figure 1). The function of these new family members is still to be elucidated but there is evidence that some may act through IL-1RL2 and IL-1RAcP14, 15 and that IL-1F7 may be important in IL-18 antagonism.16

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Graphical representation of the linkage disequilibrium (LD) patterns across the candidate loci from Haploview. Shown is the organization of the interleukin-1 (IL-1) family genes in the two clusters on chromosome 2 as well as the single nucleotide polymorphism (SNP) coverage of the initial SNP set. The LD between these SNPs is shown (r2) for those <150kb apart. Key: white square, r2=0; grey square, 0<r2<1; black square, r2=1.

Full figure and legend (317K)

The genes for the IL-1 receptor family are also located together in a cluster on chromosome 2q11.2, along with the genes for the two IL-18 receptors (IL-18R1 and IL-18RAcP) also members of the IL-1 receptor family17 (Figure 1). The type 1 IL-1 receptor (IL-1R1) is required for signal transduction, while the type 2 IL-1 receptor (IL-1R2) is a non-signal transducing, decoy receptor.18 There are also an additional two members of the IL-1 receptor family, IL-1RL1 and IL-1RL2, neither of which binds IL-119 but IL-1RL2 has been shown to be the receptor through which the IL-1F6, 8 and 9 ligands induce a signal.14

IL-1 inhibitory activity was found in the serum of sJIA patients at the height of fever20 and there has been shown to be a significant difference (10-fold increase) between the plasma levels of IL-1Ra in control individuals and in sJIA patients.21 Pascual et al.22 found that when peripheral blood mononuclear cells (PBMCs) from healthy individuals were incubated with serum from active sJIA patients several members of the IL-1 and IL-1R family: IL1B, IL1RN, IL1R1 and IL1R2, showed a more than twofold increased gene expression compared to PBMCs incubated without culture or with autologous serum. Additionally, they found that IL-1β and IL-1R2 proteins were over-expressed in the PBMCs of patients compared to healthy controls, PBMCs from sJIA patients also had a greater capacity to secrete IL-1β after stimulation with phorbol myristate acetate-ionomycin than PBMCs from healthy individuals (15-fold increase) but there was no significant difference in the production of IL-6 and TNF between patients and controls in the same cultures.22 Further evidence that IL-1 is important in the pathogenesis of sJIA comes from both papers and anecdotal evidence that treatment with a recombinant IL-1 antagonist, Anakinra, shows remarkable improvement in some sJIA patients.22 However, there is evidence that there may be at least two sJIA patient subpopulations: Anakinra-responsive and non-responsive; in a trial of French patients approximately half were responsive.23 Comparable findings were also observed in an Italian cohort and also in a UK cohort (unpublished observations). Similar results have been seen with the receptor based IL-1 blocker, IL-1 TRAP (Rilonacept).24

A number of associations of IL-1 ligand and receptor gene polymorphisms have been found with various inflammatory and autoimmune diseases, including osteoarthritis,25 severity of rheumatoid arthritis (RA)26 and the severity of chronic polyarthritis.27 There have also been a number of associations shown with ankylosing spondylitis,28, 29 susceptibility and severity of alopecia areata,30 periodontitis31 and type 1 diabetes (including circulating plasma levels in an allele dose effect).32 IL-1 family associations of particular note include a polymorphism in the IL1A promoter region with early onset oligoarticular arthritis,33 although this was not replicated in a subsequent cohort.34 A variable number tandem repeat in IL1RN has been associated with JIA as a whole, when the subtypes were examined separately this was also significant in the subtypes of extended oligoarthritis, enthesitis-related arthritis and other arthritis, but not sJIA.35

To date no IL-1 family associations have been found with the systemic subtype of JIA specifically,33, 35, 36 however, these studies were designed to investigate JIA as a whole and therefore in subgroup analyses the systemic subtype was underpowered. Previous association studies of the IL-1 family have also only looked at limited polymorphic diversity in one or two of the family members while a complete screen of all the members is necessary to be able to fully understand the involvement of IL-1 in disease susceptibility. We report here the most comprehensive gene association study to date of all the known genes within the two IL-1 gene family clusters in sJIA patients, using tagging single nucleotide polymorphisms (tSNPs).

Top

Results

Tagging SNP selection

Genotype data were publicly available for a total of 998 SNPs in the two clusters that satisfied the criteria of minor allele frequency (MAF)greater than or equal to0.05 in Caucasians. Due to the significant linkage disequilibrium (LD; Figure 1) between these SNPs, tSNP selection identified 211 SNPs capable of capturing all of the variation contained within the full set, 88 for the ligand cluster and 123 for the receptor cluster.

Genotyping

In stage 1 five SNPs in the IL1 ligand cluster and one in the IL1 receptor cluster were dropped due to low confidence in genotype calling. An additional four SNPs in the IL1 ligand cluster and six SNPs in the IL1 receptor cluster were dropped from analysis due to the controls significantly violating Hardy–Weinberg equilibrium. All the SNPs in stage 2 were in Hardy–Weinberg equilibrium and there was no significant difference found in the genotype calling of the two platforms used (97.8% concordance).

Stage 1 association analysis

In the initial individual analysis of the stage 1 data 11 SNPs in the IL1 ligand region showed significant allele frequency differences between the two populations (P<0.05). Conditional analysis showed that the observed significance of three of these SNPs was due to their being in significant LD with another of the SNPs, and so they were removed from the model. Significant allele frequency differences between the sJIA cases and the healthy controls were observed for SNP1 (rs6712572, P=0.0045), SNP4 (rs2071374, P=0.006), SNP5 (rs3783516, P=0.0053), SNP7 (rs4848123, P=0.003), SNP10 (rs3917368, P=0.0096), SNP64 (rs1688075, P=0.00089), SNP80 (rs4849159, P=0.04) and SNP81 (rs6760120, P=0.02) (Table 1).


Of the 123 tSNPs typed in the IL1 receptor region, 9 showed significant association in the initial analysis, all of which were located around IL1R2. Conditional and haplotype analysis determined that the effect shown by these nine SNPs was explained by two of them, SNP8 and SNP23 (rs12712122, P=0.003 and rs4851531, P=0.009 respectively) (Table 1). The alternative likelihoods of the analyses of this haplotype, with and without interactions, were exactly the same, showing that the best model for the association is that of no interaction, suggesting that the two markers contribute independently to disease risk but do not interact in an epistatic manner. These SNPs show odds ratios (ORs)=1.52 and 1.59 respectively, with OR=2 for the two combined.

No haplotypes examined showed stronger significance than individual SNPs and so results are not presented here.

These 10 SNPs from the two candidate regions showing evidence of association in stage 1 were pursued for further investigation in stage 2 of the study.

Stage 2 association analysis

The frequencies of the 10 SNPs typed in the stage 2 samples are shown in Table 2. None of the SNPs show a significant frequency difference in this sample set alone.


The meta-analysis of the data from the two stages of the association study identified three SNPs within the IL1 ligand cluster that showed evidence of association (P<0.05), SNP1 (P=0.025), SNP4 (P=0.002) and SNP64 (P=0.002) (Table 3). The table also shows that these three IL1 ligand cluster SNPs did not have significant (P<0.05) Breslow–Day (BD) P-values, while the four SNPs which did not show evidence of association overall did show evidence of significant differences between the ORs in the two groups tested. SNPs1 and 2 are not tagging any other SNPs within the region, but SNP64 is tagging two other SNPs: SNPs 64a (rs28928309) and 64b (rs315925), both with r2=1. SNP1 is located 22.29kb downstream of IL1A and SNP4 is within intron 4 of IL1A. SNP64a is located at position +4844 relative to the start of transcription of IL1F10, SNP64 is at position −26942 and SNP64b at −19771, both relative to the start of transcription of IL1RN (Figure 2).

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Positions of the associated single nucleotide polymorphisms (SNPs) within the interleukin-1 (IL1) ligand cluster region. The positions of the associated SNPs (indicated in black), and of the SNPs that they are tagging (indicated in grey), within the IL1 ligand cluster region are shown.

Full figure and legend (33K)


SNP8 in the IL1 receptor gene region showed evidence of association in the meta-analysis of the two stages (P=0.047; Table 3). Both of the IL1 receptor cluster SNPs tested in both stages of the study showed a significant difference between the effect sizes observed in the two groups (BD P<0.05). SNP8 is tagging four other SNPs: 8a (rs1010329), 8b (rs2190361), 8c (rs2190360) and 8d (rs740044), with r2=0.86, 1, 1 and 0.82 respectively. All five of these significant SNPs are located upstream of IL1R2 at, in chromosomal order, −43236, −37208, −16309 and −8137, relative to start of transcription (Figure 3). As SNPs 8a and 8d are not in total LD with SNP 8 they were also both subsequently typed directly. They both showed a significant frequency difference in the stage 1 samples (P=0.0062 and P=0.02), but not in the stage 2 samples (P=0.52 and P=0.81), nor in the meta-analysis of the two stages (P=0.11 and P=0.13) (data not shown).

Figure 3.
Figure 3 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Positions of the associated single nucleotide polymorphism (SNP) within the interleukin-1 (IL1) receptor cluster region. The positions of the associated SNP (indicated in black), and of the SNPs that it is tagging (indicated in grey), within the IL1 receptor cluster region are shown.

Full figure and legend (34K)

Top

Discussion

The recent publication of large genotype databases has made it possible for more comprehensive association studies to be performed by enabling investigation of the LD between SNPs in candidate genes, facilitating the use of tSNP approaches. This significantly reduces the amount of redundancy by removing polymorphisms that are in high LD with each other, and capturing the variation contained within a large number of polymorphisms using a subset. It is, therefore, now possible to carry out more comprehensive investigations of regions containing candidate genes by investigating all of the known sequence variations present, rather than, the often subjective, selection of candidate SNPs with known or predicted functionality. This approach also facilitates the discovery of novel regions that may affect gene activity and impact on disease status through any findings of disease association of SNPs within unannotated sequence.

The SNPs showing evidence of association with sJIA in the IL1 ligand cluster region are located within, or in the surrounding sequence of, IL1A, IL1F10 and IL1RN. However, as SNP64a is located 1.3kb downstream from the 3′-end of IL1F10 and 50.4kb upstream of the 5′-end of IL1RN, it is possible that it may be involved in the regulation of IL1RN and not of IL1F10. Consequently, it is possible that these associated polymorphisms are only involved with the major forms of IL-1 and that the novel, less widely expressed IL-1 family members are not involved in sJIA susceptibility. ProIL-1α is biologically active and, as well as being found on the cell membrane, is released from cells during disease, where it is a key activator of innate immunity and acute inflammatory responses. Although the effects of IL-1β are more widely investigated there is evidence that IL-1α is important in arthritis development and progression. This includes the observation that levels of membrane-bound IL-1α, but not serum IL-1α or IL-1β, correlate with the severity of arthritis in a mouse model,37 and that RA patients producing anti-IL-1α antibodies develop a less destructive disease.38 Dysregulation of the IL1A gene could lead to a perpetuation of the local inflammatory process, contributing to the development of a chronic inflammatory state. IL1RN encodes the IL-1 receptor antagonist IL-1Ra, alterations in the expression of this gene could lead to decreased levels of IL-1 antagonism, resulting in enhanced IL-1 cell signalling. If this is the case it could account for the improvement seen in some sJIA patients when treated with the recombinant IL-1 receptor antagonist Anakinra, as the balance between activation and repression of the IL-1 response would be restored closer to the non-inflammatory state.

The only SNPs in the IL1 receptor cluster that showed evidence of a significant association with disease were all in the region of IL1R2, increasing the confidence that this association is true. Although this association shows only borderline significance in the meta-analysis, we believe that this association is still interesting and warrants further investigation. SNP8 and the four SNPs it is tagging are all located upstream of IL1R2, suggesting that they are involved in regulation of transcription. As IL-1R2 has a short cytosolic domain that is non-signal transducing it is a functionally negative receptor and acts as a decoy, sequestering IL-1 and preventing it from binding to the signal transducing IL-1R1. Thus, alteration in the level of expression of IL-1R2 will in turn affect the activity level of IL-1 during inflammation. Proteolytic cleavage of the membrane bound IL-1R2, so that the extracellular domain is released into the circulation, yields a soluble form of the protein (sIL-1R2), which is found in the circulation and urine of healthy subjects, but elevated in patients with sepsis and with active RA.8 sIL-1R2 interacts with secreted IL-1RAcP to form a high-affinity IL-1 scavenger, with an approximately 100-fold increased affinity for IL-1α and IL-1β compared to sIL-1R2 alone, but no increased affinity for IL-1Ra.39 This complex binds to IL-1β with a dissociation rate of 2h,8 furthering the importance of IL-1R2 in the neutralization of IL-1. Should these associated SNPs cause a reduction in transcription then this could account for the increased levels of IL-1 and the chronic inflammation seen in active sJIA patients.

The SNPs used in this study are acting as tSNPs for other polymorphisms within the regions, therefore those showing evidence of association are not necessarily directly responsible for the effect seen, rather a SNP being tagged could be the functional polymorphism affecting disease susceptibility. Although the initial SNP coverage was as comprehensive as possible at the time, it did not include the total genetic diversity across the candidate loci; therefore it is possible that the SNPs identified in this study as being associated with sJIA are in LD with an unknown, hidden, functional polymorphism responsible for the effect seen.

One potential limitation of this study is the small sample size, and therefore its low power to detect associations. As a result it is possible that there are other associated SNPs with smaller effect that were not detected in this study. sJIA is a rare disease with a prevalence of approximately only 16 children in 100000, and only approximately 20 new patients being seen a year at Great Ormond Street Hospital (GOSH). For these reasons larger patient cohort numbers are extremely difficult to collect. This study has already utilized a UK repository of sJIA patients, further to which it would not be possible to collect a sufficiently large number of additional patients to significantly increase the power of this study. However, previous case–control studies investigating sJIA have used similar, and smaller, sample sizes, for example 92 patients,3 the results of which were confirmed in a larger family study,40 so this study is favourably comparable to previous work done on this disease. The two-staged meta-analysis approach used in this study is one that minimizes population stratification in small samples, and therefore the SNPs that are found to be associated with sJIA after this analysis are more likely to be true disease susceptibility/severity genes in sJIA. Comparison of the genotype frequencies in the patient populations from this study with the frequencies in the control population from the Welcome Trust Case Control Consortium41 (available as direct genotyping data for SNPs 5 and 81, and as imputed genotyping data42 for all others, except SNP4 that is not represented) shows the same pattern of significant differences as reported here, that is significant frequency differences observed with the stage 1 patients, and with all the patients pooled together, but not with the patients from stage 2 alone. The fact that similar results are observed using this larger control cohort (n=2936), which provides greater sensitivity, indicates that the associations found in this study are most likely to be real. To confirm the findings of this study, replication using a larger, independent patient cohort is necessary.

A potential problem created by this high-density marker approach is that of multiple testing; a necessity in studies with a low prior probability. However, it is widely accepted that correcting for multiple testing using traditional methods such as Bonferroni correction is not appropriate in studies such as this, as it assumes that all of the variables being tested are independent. The result is that correction of the P-values would be overly conservative since the SNPs are in some amount of LD with each other.43 Other correction methods such as Bayesian inference and false-positive reporting probability also pose difficulties as they depend on prior probability, appropriate values for which are impossible to determine when neither the number of genes involved in a complex disease, nor their effect sizes, is known. This issue must therefore be taken into consideration when interpreting these results, and replication of the findings is essential before any definite conclusions can be drawn.

The observation that just under half of patients treated with Anakinra show a positive response23 suggests that there may be at least two different populations of sJIA patients; in only some of whom the main underlying cause for disease is a defect in IL-1 regulation. This means that the associations observed in this study may only have a significant effect in those patients responsive to Anakinra, potentially confounding the results due to admixture. It is however not possible to stratify the data in this study by Anakinra response as currently it is only prescribed to patients who are refractory to the standard anti-inflammatory drugs, and so only a very small proportion of the patients used in this study have been treated with Anakinra. A study to investigate the effects of these associated SNPs in patients responsive and refractory to Anakinra treatment is currently being planned.

If there is a higher proportion of patients who are IL-1-blockade responsive in one group than in the other it may explain why six of the ten SNPs typed in both stages of this study had significant BD P-values; showing that there is a significant difference in the effect sizes in the two groups tested. The majority of the SNPs in this study showing evidence of heterogeneity of ORs are not significantly associated with disease status; these negative findings could potentially be positive associations when stratified by Anakinra response, or it is possible that the frequency differences for these SNPs observed in stage 1 were spurious false-positive findings, abrogated when the additional samples from stage 2 are included in the analysis.

In conclusion, this study has found a number of SNPs in IL1R2, IL1A, IL1F10 and IL1RN to show evidence of association with sJIA. This is the first study to show genetic associations between members of the IL-1 gene family and the systemic subtype of JIA, indicating that the IL-1 family is important in the development of disease severity and susceptibility. These finding give a better understanding of the underlying aetiological cause of systemic JIA and could lead to new diagnostic, pharmacogenetic and therapeutic targets to be developed.

Top

Materials and methods

Selection of tSNPs for typing

Genotyping data for all the SNPs within the candidate regions were obtained from the International HapMap project44 for the Centre d'Etude du Polymorphisme Humain (CEPH; Utah residents with ancestry from northern and western Europe) population (HapMap data release no. 18/phase II Sept. 05, on NCBI B34 assembly, dbSNPb124). These regions covered the gene and the surrounding sequences both upstream and downstream to the next flanking genes. In addition, genotyping data for the population of European descent were obtained from the Programmes for Genomic Applications (www.nhlbi.nih.gov/resources/pga) for all of the candidate genes that they had re-sequenced (all candidate genes except IL1RL1 and IL18R2). As there was overlap in the individuals typed by the two data sources all of the genotyping data were combined into one data file with more complete SNP coverage.

The LD between each of these SNPs that had an MAF of at least 0.05 was calculated using the Haploview programme (version 3.2).45 The LD data were then used in the Tagger program46 within Haploview to select tSNPs, examining for pairwise tagging with a minimum r2 value of 0.8.

Study strategy

To maximize the power of the study given the limited resources available a two-stage study design47, 48 was adopted. In this strategy a group of patient and control samples were typed for all of the SNPs of interest in stage 1, and another group of patients and controls from a different source were typed in stage 2 only for those SNPs that showed evidence of association in stage 1. Due to the samples used in the two stages having been collected from different sources they cannot be combined into a single analysis; therefore the results from both stages of the study were analysed together by meta-analysis, as discussed below.

The power of this meta-analysis strategy is difficult to estimate. As a guide, a two-stage study design, given the number, and proportion of samples and SNPs genotyped in each stage, gives 84% power to detect an SNP with a frequency of 0.3 giving a relative risk of 1.7 using a multiplicative model (CaTS power calculator for two-stage association analysis).48

Patient and control samples

DNA from Caucasian sJIA patients, collected from the British Society for Paediatric and Adolescent Rheumatology Study Group National DNA repository at the Arthritis Research Campaign Epidemiology Unit in Manchester and the University College London Centre for Paediatric and Adolescent Rheumatology at both GOSH and the Middlesex Hospital, was used in stage 1 (n=130). DNA from further Caucasian patients from GOSH in London, Wexham Park Hospital, Berkshire, United Kingdom, and Necker Hospital in Paris was used for stage 2 (n=105). Ethical approval for the study was obtained (Great Ormond Street Hospital for Children NHS Trust and Institute of Child Health, Research Ethics Committee reference 02RU06) and parents gave informed consent. Control DNA used in stage 1 was collected from an ethnically matched population of 16 to 30-year olds from a GP practice in a stable population of the west Midlands (n=151), and that used in stage 2 from Caucasian first-time blood donors from the national blood transfusion centre in London (n=184).

Genotyping

Genotyping for stage 1 of the study was done with the Illumina Golden Gate assay genotyping platform (custom 384 SNP panel, 96-sample Sentrix array matrix). Automatic clustering of the samples and genotype calling was done with the Illumina BeadStudio software (version 2.1.8.32932). Determination of the validity of the genotyping of each SNP was based on both the software-generated Gen Train score (based on the shapes of the clusters and the relative distance between them) and visual examination of the genotype clustering; any SNPs without three easily identifiable, compact and well-spaced clusters were considered to have failed genotyping. All clustering and genotype calling was confirmed independently by two investigators. Genotyping for stage 2 of the study was performed by KBioscience using Applied Biosystems TaqMan assays. A proportion (18) of the sJIA samples used in the initial study were also re-typed to confirm there were no discrepancies between the two genotyping methods.

Genotype data checking

The genotypes for the control populations were checked for Hardy–Weinberg equilibrium using a χ2-test with a significant P-value cut-off of 0.05. To check that the control population used for genotyping and the CEPH population used to select the tSNPs were comparable, the distribution of the populations between the three possible genotypes for each SNP was compared using a χ2 contingency table with a significant P-value cut-off of 0.05.

Stage 1 association analysis

Association analysis of the genotyping data was performed using the COCAPHASE module of UNPHASED (v2.403 and v3.0), which performs unconditional logistic regression likelihood ratio tests under a log-linear model.49 Both single marker and haplotype analyses were performed and haplotype frequencies were estimated by an expectation-maximization algorithm in COCAPHASE. When multiple SNPs in the same region showed a significant difference they were separated into clusters according to their LD, both D′ and r2, patterns. A stepwise conditional regression approach50 was then used within each LD cluster to identify the SNPs showing primary effects. The stepwise conditional regression was then repeated with the primary effect SNPS from each LD cluster to determine if their effects were independent or related to each other. Nested conditional analysis was also performed to ensure that all of the primary SNPs were required to fully explain the effects seen.

Stage 2 association analysis

The genotyping data from the two stages of the association study were analysed for association using the Cochran–Mantel–Haenszel test for meta-analysis, and the BD test for homogeneity of ORs between the two stages, using the software PLINK.51

Top

References

  1. Petty RE. Growing pains: the ILAR classification of juvenile idiopathic arthritis. J Rheumatol 2001; 28: 927–928. | PubMed | ChemPort |
  2. Woo P. Systemic juvenile idiopathic arthritis: diagnosis, management, and outcome. Nat Clin Pract Rheumatol 2006; 2: 28–34. | Article | PubMed |
  3. Fishman D, Faulds G, Jeffery R, Mohamed-Ali V, Yudkin JS, Humphries S et al. The effect of novel polymorphisms in the interleukin-6 (IL-6) gene on IL-6 transcription and plasma IL-6 levels, and an association with systemic-onset juvenile chronic arthritis. J Clin Invest 1998; 102: 1369–1376. | Article | PubMed | ISI | ChemPort |
  4. Ogilvie EM, Fife MS, Thompson SD, Twine N, Tsoras M, Moroldo M et al. The - 174G allele of the interleukin-6 gene confers susceptibility to systemic arthritis in children: a multicenter study using simplex and multiplex juvenile idiopathic arthritis families. Arthritis Rheum 2003; 48: 3202–3206. | Article | PubMed | ISI | ChemPort |
  5. Donn RP, Shelley E, Ollier WE, Thomson W. A novel 5'-flanking region polymorphism of macrophage migration inhibitory factor is associated with systemic-onset juvenile idiopathic arthritis. Arthritis Rheum 2001; 44: 1782–1785. | Article | PubMed | ISI | ChemPort |
  6. de Benedetti F, Meazza C, Vivarelli M, Rossi F, Pistorio A, Lamb R et al. Functional and prognostic relevance of the - 173 polymorphism of the macrophage migration inhibitory factor gene in systemic-onset juvenile idiopathic arthritis. Arthritis Rheum 2003; 48: 1398–1407. | Article | PubMed | ISI | ChemPort |
  7. Fife MS, Gutierrez A, Ogilvie EM, Stock CJ, Samuel JM, Thomson W et al. Novel IL10 gene family associations with systemic juvenile idiopathic arthritis. Arthritis Res Ther 2006; 8: R148. | Article | PubMed | ChemPort |
  8. Dinarello CA. Biologic basis for interleukin-1 in disease. Blood 1996; 87: 2095–2147. | PubMed | ISI | ChemPort |
  9. Nicklin MJ, Barton JL, Nguyen M, FitzGerald MG, Duff GW, Kornman K. A sequence-based map of the nine genes of the human interleukin-1 cluster. Genomics 2002; 79: 718–725. | Article | PubMed | ChemPort |
  10. Taylor SL, Renshaw BR, Garka KE, Smith DE, Sims JE. Genomic organization of the interleukin-1 locus. Genomics 2002; 79: 726–733. | Article | PubMed | ChemPort |
  11. Smith DE, Renshaw BR, Ketchem RR, Kubin M, Garka KE, Sims JE. Four new members expand the interleukin-1 superfamily. J Biol Chem 2000; 275: 1169–1175. | Article | PubMed | ISI | ChemPort |
  12. Kumar S, McDonnell PC, Lehr R, Tierney L, Tzimas MN, Griswold DE et al. Identification and initial characterization of four novel members of the interleukin-1 family. J Biol Chem 2000; 275: 10308–10314. | Article | PubMed | ISI | ChemPort |
  13. Lin H, Ho AS, Haley-Vicente D, Zhang J, Bernal-Fussell J, Pace AM et al. Cloning and characterization of IL-1HY2, a novel interleukin-1 family member. J Biol Chem 2001; 276: 20597–20602. | Article | PubMed | ChemPort |
  14. Debets R, Timans JC, Homey B, Zurawski S, Sana TR, Lo S et al. Two novel IL-1 family members, IL-1 delta and IL-1 epsilon, function as an antagonist and agonist of NF-kappa B activation through the orphan IL-1 receptor-related protein 2. J Immunol 2001; 167: 1440–1446. | PubMed | ISI | ChemPort |
  15. Towne JE, Garka KE, Renshaw BR, Virca GD, Sims JE. Interleukin (IL)-1F6, IL-1F8, and IL-1F9 signal through IL-1Rrp2 and IL-1RAcP to activate the pathway leading to NF-kappaB and MAPKs. J Biol Chem 2004; 279: 13677–13688. | Article | PubMed | ISI | ChemPort |
  16. Bufler P, Azam T, Gamboni-Robertson F, Reznikov LL, Kumar S, Dinarello CA et al. A complex of the IL-1 homologue IL-1F7b and IL-18-binding protein reduces IL-18 activity. Proc Natl Acad Sci USA 2002; 99: 13723–13728. | Article | PubMed | ChemPort |
  17. Dale M, Nicklin MJ. Interleukin-1 receptor cluster: gene organization of IL1R2, IL1R1, IL1RL2 (IL-1Rrp2), IL1RL1 (T1/ST2), and IL18R1 (IL-1Rrp) on human chromosome 2q. Genomics 1999; 57: 177–179. | Article | PubMed | ChemPort |
  18. Stylianou E, O'Neill LA, Rawlinson L, Edbrooke MR, Woo P, Saklatvala J. Interleukin 1 induces NF-kappa B through its type I but not its type II receptor in lymphocytes. J Biol Chem 1992; 267: 15836–15841. | PubMed | ISI | ChemPort |
  19. Subramaniam S, Stansberg C, Cunningham C. The interleukin 1 receptor family. Dev Comp Immunol 2004; 28: 415–428. | Article | PubMed | ISI | ChemPort |
  20. Prieur AM, Roux-Lombard P, Dayer JM. Dynamics of fever and the cytokine network in systemic juvenile arthritis. Rev Rhum Engl Ed 1996; 63: 163–170. | PubMed | ChemPort |
  21. Muller K, Herner EB, Stagg A, Bendtzen K, Woo P. Inflammatory cytokines and cytokine antagonists in whole blood cultures of patients with systemic juvenile chronic arthritis. Br J Rheumatol 1998; 37: 562–569. | Article | PubMed | ChemPort |
  22. Pascual V, Allantaz F, Arce E, Punaro M, Banchereau J. Role of interleukin-1 (IL-1) in the pathogenesis of systemic onset juvenile idiopathic arthritis and clinical response to IL-1 blockade. J Exp Med 2005; 201: 1479–1486. | Article | PubMed | ISI | ChemPort |
  23. Lequerre T, Quartier P, Rosellini D, Alaoui F, De Bandt M, Mejjad O et al. Interleukin-1 receptor antagonist (anakinra) treatment in patients with systemic-onset juvenile idiopathic arthritis or adult onset Still's disease. Preliminary experience in France. Ann Rheum Dis 2008; 67: 302–308. | Article | PubMed | ChemPort |
  24. Lovell D, Giannini E, Kinura Y, Li S, Hashkes P, Reiff A et al. Preliminary evidence for bioactivity of IL-1 TRAP (Rilonacept), a long acting IL-1 inhibitor, in systemic juvenile idiopathic arthritis. Ann Rheum Dis 2006; 65 (Suppl II): 693.Ref Type: Abstract.
  25. Meulenbelt I, Seymour AB, Nieuwland M, Huizinga TW, van Duijn CM, Slagboom PE. Association of the interleukin-1 gene cluster with radiographic signs of osteoarthritis of the hip. Arthritis Rheum 2004; 50: 1179–1186. | Article | PubMed | ChemPort |
  26. Buchs N, di Giovine FS, Silvestri T, Vannier E, Duff GW, Miossec P. IL-1B and IL-1Ra gene polymorphisms and disease severity in rheumatoid arthritis: interaction with their plasma levels. Genes Immun 2001; 2: 222–228. | Article | PubMed | ChemPort |
  27. Jouvenne P, Chaudhary A, Buchs N, Giovine FS, Duff GW, Miossec P. Possible genetic association between interleukin-1alpha gene polymorphism and the severity of chronic polyarthritis. Eur Cytokine Netw 1999; 10: 33–36. | PubMed | ChemPort |
  28. Timms AE, Crane AM, Sims AM, Cordell HJ, Bradbury LA, Abbott A et al. The interleukin 1 gene cluster contains a major susceptibility locus for ankylosing spondylitis. Am J Hum Genet 2004; 75: 587–595. | Article | PubMed | ISI | ChemPort |
  29. van der PM, Crusius JB, Garcia-Gonzalez MA, Baudoin P, Kostense PJ, Alizadeh BZ et al. Interleukin-1beta and interleukin-1 receptor antagonist gene polymorphisms in ankylosing spondylitis. Rheumatology (Oxford) 2002; 41: 1419–1423. | Article | PubMed |
  30. Tazi-Ahnini R, Cox A, McDonagh AJ, Nicklin MJ, di Giovine FS, Timms JM et al. Genetic analysis of the interleukin-1 receptor antagonist and its homologue IL-1L1 in alopecia areata: strong severity association and possible gene interaction. Eur J Immunogenet 2002; 29: 25–30. | Article | PubMed | ISI | ChemPort |
  31. Kinane DF, Hart TC. Genes and gene polymorphisms associated with periodontal disease. Crit Rev Oral Biol Med 2003; 14: 430–449. | PubMed | ChemPort |
  32. Bergholdt R, Larsen ZM, Andersen NA, Johannesen J, Kristiansen OP, Mandrup-Poulsen T et al. Characterization of new polymorphisms in the 5' UTR of the human interleukin-1 receptor type 1 (IL1R1) gene: linkage to type 1 diabetes and correlation to IL-1RI plasma level. Genes Immun 2000; 1: 495–500. | Article | PubMed | ChemPort |
  33. McDowell TL, Symons JA, Ploski R, Forre O, Duff GW. A genetic association between juvenile rheumatoid arthritis and a novel interleukin-1 alpha polymorphism. Arthritis Rheum 1995; 38: 221–228. | Article | PubMed | ISI | ChemPort |
  34. Donn RP, Farhan AJ, Barrett JH, Thomson W, Worthington J, Ollier WE. Absence of association between interleukin 1 alpha and oligoarticular juvenile chronic arthritis in UK patients. Rheumatology (Oxford) 1999; 38: 171–175. | Article | PubMed | ChemPort |
  35. Vencovsky J, Jarosova K, Ruzickova S, Nemcova D, Niederlova J, Ozen S et al. Higher frequency of allele 2 of the interleukin-1 receptor antagonist gene in patients with juvenile idiopathic arthritis. Arthritis Rheum 2001; 44: 2387–2391. | Article | PubMed | ChemPort |
  36. Sugiura T, Maeno N, Kawaguchi Y, Takei S, Imanaka H, Kawano Y et al. A promoter haplotype of the interleukin-18 gene is associated with juvenile idiopathic arthritis in the Japanese population. Arthritis Res Ther 2006; 8: R60. | Article | PubMed | ChemPort |
  37. Niki Y, Yamada H, Kikuchi T, Toyama Y, Matsumoto H, Fujikawa K et al. Membrane-associated IL-1 contributes to chronic synovitis and cartilage destruction in human IL-1 alpha transgenic mice. J Immunol 2004; 172: 577–584. | PubMed | ChemPort |
  38. Graudal NA, Svenson M, Tarp U, Garred P, Jurik AG, Bendtzen K. Autoantibodies against interleukin 1alpha in rheumatoid arthritis: association with long term radiographic outcome. Ann Rheum Dis 2002; 61: 598–602. | Article | PubMed | ChemPort |
  39. Smith DE, Hanna R, Della F, Moore H, Chen H, Farese AM et al. The soluble form of IL-1 receptor accessory protein enhances the ability of soluble type II IL-1 receptor to inhibit IL-1 action. Immunity 2003; 18: 87–96. | Article | PubMed | ChemPort |
  40. Ogilvie EM, Fife MS, Thompson SD, Twine N, Tsoras M, Moroldo M et al. The - 174G allele of the interleukin-6 gene confers susceptibility to systemic arthritis in children: a multicenter study using simplex and multiplex juvenile idiopathic arthritis families. Arthritis Rheum 2003; 48: 3202–3206. | Article | PubMed | ISI | ChemPort |
  41. Wellcome Trust Case Control Consortium. Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature 2007; 447: 661–678. | Article | PubMed | ISI | ChemPort |
  42. Marchini J, Howie B, Myers S, McVean G, Donnelly P. A new multipoint method for genome-wide association studies by imputation of genotypes. Nat Genet 2007; 39: 906–913. | Article | PubMed | ISI | ChemPort |
  43. Lewis CM. Genetic association studies: design, analysis and interpretation. Brief Bioinform 2002; 3: 146–153. | Article | PubMed | ChemPort |
  44. The International HapMap Consortium. The International HapMap Project. Nature 2003; 426: 789–796. | Article | PubMed | ISI | ChemPort |
  45. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005; 21: 263–265. | Article | PubMed | ISI | ChemPort |
  46. de Bakker PI, Yelensky R, Pe'er I, Gabriel SB, Daly MJ, Altshuler D. Efficiency and power in genetic association studies. Nat Genet 2005; 37: 1217–1223. | Article | PubMed | ISI | ChemPort |
  47. Satagopan JM, Venkatraman ES, Begg CB. Two-stage designs for gene-disease association studies with sample size constraints. Biometrics 2004; 60: 589–597. | Article | PubMed | ISI |
  48. Skol AD, Scott LJ, Abecasis GR, Boehnke M. Joint analysis is more efficient than replication-based analysis for two-stage genome-wide association studies. Nat Genet 2006; 38: 209–213. | Article | PubMed | ISI | ChemPort |
  49. Dudbridge F. Pedigree disequilibrium tests for multilocus haplotypes. Genet Epidemiol 2003; 25: 115–121. | Article | PubMed | ISI |
  50. Cordell HJ, Clayton DG. A unified stepwise regression procedure for evaluating the relative effects of polymorphisms within a gene using case/control or family data: application to HLA in type 1 diabetes. Am J Hum Genet 2002; 70: 124–141. | Article | PubMed | ISI | ChemPort |
  51. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007; 81: 559–575. | Article | PubMed | ISI | ChemPort |
Top

Acknowledgements

We thank BSPAR study group members: Dr M Abinun, Dr M Becker, Dr A Bell, Professor A Craft, Dr E Crawley, Dr J David, Dr H Foster, Dr J Gardener-Medwin, Dr J Griffin, Dr A Hall, Dr M Hall, Dr A Herrick, Dr P Hollingworth, Dr L Holt, Dr S Jones, Dr G Pountain, Dr C Ryder, Professor T Southwood, Dr I Stewart, Dr H Venning, Dr L Wedderburn, Professor P Woo and Dr S Wyatt; Dr AM Prieur and the French association Kourir; Dr J Packham, for the contribution of patient DNA and Dr Ele Zeggini for assistance with analysis. This work was funded by the Nuffield Foundation Oliver Bird Rheumatism Program and the Arthritis Research Campaign (arc).

Top

MORE ARTICLES LIKE THIS

These links to content published by NPG are automatically generated